Model.join()

Contents

Model.join()#

Model.join(model, by=None, copy=False, suffix=('.x', '.y'))[source]#

Joins a second model into this one.

Parameters:

  • model (Model) – The second model to join into this one.

  • by (Sequence[str] | None, default: None) – Sequence of variable names to join on.

  • copy (bool, default: False) – Whether to copy the second model before joining.

  • suffix (tuple[str, str], default: ('.x', '.y')) – Suffixes to use for renaming of variables with duplicate names.

See also

Model.join_by_all

Automatically join by all overlapping names.

Return type:

Self

Notes

If there are variables with duplicate names, the method’s behavior depends on by:

  1. If the duplicate name is supplied in by, then the variables from self (i.e., the model on which the method is called) are used. They replace the respective variables in the second model.

  2. If the duplicate name is not supplied in by, then the duplicate names are resolved by renaming the respective variables from both models using suffix.

The seed nodes of the second model are added to the calling model’s seed nodes.

Examples

Nothing supplied in by, duplicate names are resolved by renaming:

>>> import liesel.model as lsl
>>> x1 = lsl.Var.new_obs(1.0, name="x")
>>> x2 = lsl.Var.new_obs(1.0, name="x")
>>> y = lsl.Var.new_calc(lambda x: x, x2, name="y")
>>> m1 = lsl.Model(x1)
>>> m2 = lsl.Model(x2, y)
>>> m1.join(m2)
Model(9 nodes, 3 vars)
>>> list(m1.vars)
['x.y', 'x.x', 'y']
>>> list(m2.vars)
[]
>>> m1.seed_nodes_and_vars
[Var(name="x.x"), Var(name="x.y"), Var(name="y")]

Joining on ‘x’:

>>> import liesel.model as lsl
>>> x1 = lsl.Var.new_obs(1.0, name="x")
>>> x2 = lsl.Var.new_obs(1.0, name="x")
>>> y = lsl.Var.new_calc(lambda x: x, x2, name="y")
>>> m1 = lsl.Model(x1)
>>> m2 = lsl.Model(x2, y)
>>> m1.join(m2, by=["x"])
Model(7 nodes, 2 vars)
>>> list(m1.vars)
['x', 'y']
>>> list(m2.vars)
[]
>>> y.value_node[0] is x1
True
>>> m1.seed_nodes_and_vars
[Var(name="x"), Var(name="y")]
Contents